Edit was because it’s uncalled for. Hoped you wouldn’t have to read it. Sorry about that.

The article makes a presumption, that the active listening is actually sending voice data as audio. Then tries to splurdge, that “acsthually it’s other data”

Then tries to splurdge, that they would require to download all “wanted” words as keywords, and it wouldn’t be feasable.

Not like you would only need some words of intent “I would like to (enter 10 s of transcription)” and just hit send.

The whole article smells of washing, and the question is directed to other people, who maybe followed the story more closely, and actually has the idea what exactly is “active listening”. Maybe someone reversed engineered it.

Thanks for your useless comment

Wasn’t the whole CMG active listening about sending transcription data only? It’s not like it can’t be done directly on the device…